Mythos class for everyone*

From doomsday machines to utopian engines

For most of this decade the standard nightmare had a specific floor plan: models powerful enough to reorder the economy, locked inside two or three labs, rented out by the token, aligned to a terms-of-service you didn't write. The doomsday-machine framing wasn't about the models doing evil — it was about concentration: a handful of orgs holding capability everyone else could only lease.

That's why this month feels like a hinge. The Mythos class — Anthropic's own tiering for what sits above the classical frontier — turned out not to stay put. Fable 5 brought Mythos-grade orchestration to anyone with a Cowork subscription and a laptop; the promise of "AI for everyone" finally has a consumer-shaped front door. And now Kimi K3 arrives from the other direction: not gated, not leased — published. Weights on Hugging Face by July 27. roon, an OpenAI researcher, said it plainly: "the era of the chinese labs being far behind is over… people have to think differently now without any competitive margin built in."

What "capable" concretely means now

Utopian essays usually skip this part, so let's not: the scenarios below are only as real as the measured capability behind them. Here's what an open(-promised) model demonstrably does today — full workup in part 1: 57.1 on Artificial Analysis's Intelligence Index against Fable's 59.9. 85% on Terminal-Bench 2.1, independently measured. #1 on Arena's blind frontend-code arena — above Fable, above Sol, judged by developers who didn't know which model they were rating. $0.94 per long-horizon task. A million tokens of context. "Insanely good", per an MIT/DeepMind researcher, and "impossible to explain through distillation alone."

Translation table, benchmarks → life:

• Terminal-Bench 85% means a two-person NGO in Nairobi gets a sysadmin that never sleeps, for the cost of a used GPU server — patching, migrating, debugging the donation platform at 3am.
• Arena #1 on frontend means the village clinic's patient-intake system no longer waits for a donor-funded dev shop. Describe it, get it, own it — no per-seat license, no vendor sunset.
• 1M context + $0.30/M cached input means a solo public defender loads the entire case file, every precedent, every transcript, and asks questions all afternoon for less than lunch.
• Open weights mean a hospital in a country under sanctions, a research lab with no dollar account, a school district with a privacy mandate — all run it on-prem, forever, and nobody can switch it off remotely. That last property cuts both ways; that's part 3.

None of this requires K3 to be the best model on earth. It requires it to be good enough, ownable, and cheap — and for the first time all three are true at 2.8 points off the frontier.

The witnesses

The reaction to K3 splits along a revealing line: the people whose business model it threatens, and everyone else. Clem Delangue (Hugging Face): "the countries or companies that are leading in open science and open source AI will start leading the frontier a few years later… That's how the US took the lead." Emad Mostaque: "a true frontier model, open source! Congrats!" — with a training-cost estimate of $15–25M, pocket change against frontier training budgets. Investor Gavin Baker called it "potentially negative for Anthropic and OpenAI while being net positive for essentially every other company in the world" — which is roughly the definition of democratization written in sell-side prose. And some people didn't wait for the discourse: Aditya Agarwal — "I am literally switching models off of Fable right now for our systems… why would you pay the price if there is a good and free alternative?"

"Potentially negative for Anthropic and OpenAI while being net positive for essentially every other company in the world" is what democratization looks like when a fund manager says it.

The IF that carries the whole sentence

The Fable-plus-Cowork promise — Mythos-class intelligence for everyone — has always ended with a conditional clause: if everyone can afford it. K3 relocates that clause; it doesn't delete it.

Run the numbers from part 1 again, coldly. Self-hosting wants 64+ accelerator supernodes — Moonshot's own deployment guidance. That's not a Raspberry Pi in a community center; that's a capex line only a government, university, or mid-size company clears. The API is cheap per token and hungry in tokens: 130M output tokens to finish AA's index, double the peer median, at 62 tok/s. Fortune notes it's the most expensive model any Chinese lab has ever shipped. And Dean Ball's hands-on caveat — "very token hungry… I doubt it's actually that cheap to run" — is the sound of the fine print being read aloud.

So the honest version of the utopia is tiered, the way electrification was: first the cities, then the towns, then — decades later, after the infrastructure caught up — everyone. Weights-on-Hugging-Face is the transmission line reaching your county. The bill still exists; hosting co-ops, quantized community builds, national compute programs are the rural cooperatives of this decade, and they're already forming. The gap between "technically anyone can run it" and "practically everyone benefits" is where the next five years of this story happens.

What to actually do

• If you build for people who couldn't pay frontier prices: your cost floor just collapsed. Prototype on the K3 API now; plan for self-host or a hosting co-op when weights land July 27.
• If you're all-in on closed frontier: keep the Mythos-class model where it earns its margin — orchestration, judgment, the last 2.8 points — and route the rest downmarket. Paying Fable prices for K3-shaped work is now a choice, not a necessity.
• If you're waiting for the utopia: it arrives as infrastructure, not as a moment. Watch three gauges: the gap line above, the price of a self-host rig, and how many hosting co-ops exist by December.
• And read part 3 — because every property that makes this beautiful (ownable, un-switch-off-able, free) is also the exact property that should scare you.

The doomsday-machine era assumed the power would stay concentrated. The utopian-engine era assumes it diffuses safely. Both assumptions are currently being tested live, on all of us, with no control group. The gap to the frontier is 2.8 points and closing. Up next: what happens if the optimists are wrong.